基于Hadoop的协同过滤推荐并行化研究

Research on parallelization of collaborative filtering recommendation based on Hadoop

针对协同过滤(CF)推荐技术处理大数据时的计算效率问题,分析了CF算法的并行化。并行化CF算法采用Hadoop平台的Map Reduce并行编程模型,改善大数据环境下CF算法在单机运行时的计算性能。在实验部分,设计不同集群环境下的加速比实验,验证该算法在大数据环境中具有的计算性能。

For the computational efficiency problem existing in big data processing with collaborative filtering（CF）recommendation, parallel computing of CF is analyzed. Parallelized CF algorithm uses Map Reduce parallel programming model on Hadoop platform, which improves the computational efficiency of single PC to process big data. In the experiment section, the speedup experiments in different cluster environments are designed to verify the better computing performance of the algorithm in big data processing.